INDEX

Explanations

statements made by observers or writers reflecting on various issues or topics

oai_token-act-pair · gpt-3.5-turbo

verbs that indicate observation or commentary

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rone

-0.77

owers

-0.71

cele

-0.70

Cele

-0.69

offic

-0.68

ratch

-0.68

ugal

-0.68

onder

-0.67

productive

-0.66

inal

-0.64

POSITIVE LOGITS

ometimes

0.71

olate

0.68

omething

0.66

 petertodd

0.65

 herself

0.64

:]

0.64

 Solution

0.63

ilver

0.63

 Parables

0.63

olation

0.62

Activations Density 0.249%

statements made by observers or writers reflecting on various issues or topics

verbs that indicate observation or commentary

No Comments

No Known Activations